I read somewhere that during RMAN backups, RMAN reads the blocks for backup and skips the blocks which are being updated, and revisits at a later time to copy the blocks. Suppose I am trying to copy a huge database online, which is a 24/7 busy database with lot of updates. How is RMAN going to be a faster backup solution? Could you please explain how RMAN does internally while backing up the database?

    Requires Free Membership to View

    Login

What you read is not my experience with RMAN. As far as I know, RMAN does not skip blocks being updated. In fact, when a record is being updated, the database generally does not update the block on disk (which RMAN is reading) until some time after the commit. There are exceptions to this generalization though. RMAN will backup a block that is being changed. To ensure data consistency, RMAN uses the archived redo logs to recreate a consistent block.

This was first published in January 2008

Join the conversationComment

Share
Comments

    Results

    Contribute to the conversation

    All fields are required. Comments will appear at the bottom of the article.
    Back to top